TESS variability of B-type supergiants in the Galaxy

TESS variability of B-type supergiants in the Galaxy Michalis Kourniotis1, Lydia S. Cidale2,3, Michaela Kraus1 Matias A. Ruiz Diaz2,3, and Aldana Alberici Adam2,3 1Astronomical Institute, Czech Academy of Sciences, Fričova 298, 251 65 Ondřejov, Czech Republic, 2 Instituto de Astrofísica La Plata, CCT La Plata, CONICET-UNLP, Paseo del Bosque S/N, B1900FWA La Plata, Argentina, 3 Departamento de Espectroscopía, Facultad de Ciencias Astronómicas y Geofísicas, Universidad Nacional de La Plata (UNLP), Paseo del Bosque S/N, B1900FWA, La Plata, Argentina. Motivation : The class of blue supergiants (BSGs) consists of objects emerging from diverse evolutionary channels, which span an appreciable fraction of the massive star lifetime between the main sequence and the pre-supernova stages. The observed population statistics of BSGs contradict the expectations of the standard evolutionary theory, highlighting the need to revisit the origin and evolutionary state of these objects. At the same time, the physics that govern the nature of BSGs are imprinted into their variability profile, which provides powerful diagnostics of both the processes in the stellar interior and the dynamics of the atmosphere and wind. Motivated by this potential, we conducted and present a variability study of Galactic B-type BSGs with known spectroscopic parameters, using time-series photometry from the Transiting Exoplanet Survey Satellite (TESS). By linking the variability properties to the fundamental stellar parameters, we aim to advance our understanding of the post-main-sequence evolution, to place tight constraints on the models of the interior structure, and to distinguish between the different evolutionary pathways. Funded by the European Union (Project 101183150 - OCEANS) and the Czech Science Foundation (GAČR grant number 25-17532S). LC acknowledges financial support from CONICET (PIP 1337) and the University of La Plata (Programa de Incentivos 11/G160). The Astronomical Institute Ondřejov is supported by RVO:67985815. MRD and AAA acknowledge support from a CONICET fellowship. Fig. 1. Time-series photometry from TESS of selected BSGs. For each star, we show data from different sectors over multiple panels. Consecutive sectors are joined together, and their numbers are displayed at the bottom of each panel. Fig. 3. Stellar parameters, log(L/L⊙ ), log(Teff/K) and Eddington factor Γε of the BSGs, against their time-domain statistics: photometrc dispersion (left), coherency parameter (middle), and skewness (right). The red points indicate the (cand.) rotating variables, whereas the outer circles denote variables of the α Cygni class. We mark stars susceptible to TESS contamination with “X” symbols. Fig. 2. Frequency spectra of selected BSGs, calculated at the beginning (cyan) and the end (red) of the pre-whitening process. The vertical lines at the top of each panel point to the different types of identified frequencies: independent (solid), harmonics (dashed), and combinations (dotted). The solid black line corresponds to the best-fit SLF variability model. The gray-shaded region indicates the range of frequencies that can be identified as being due to the stellar rotation. The data & methodology Fig. 4. Hertzsprung-Russell diagram for the evolution of massive stars. Theoretical tracks at solar metallicity are taken from Ekström et al. (2012) for stars rotating initially at 40% of their critical velocity. The size of the markers is proportional to the amplitude α0 of the SLF variability. The α Cygni variables are indicated by outer circles, and the thick line represents the Humphreys-Davidson limit. The uncertainty in the temperature is illustrated by the error bar on the upper right. Fig. 5. Absolute G−band magnitudes vs. log α0. The markers are color coded as a function of logTeff. The designation for the α Cygni variables and for stars susceptible to TESS contamination follows that in Fig. 3. We show the linear fit to BSGs with MG fainter than –6.4 mag (thick solid line). The fit (+offset) to the respective parameters of BSGs in the LMC is displayed (dashed line; Bowman et al. 2019a). Results Discussion – Concluding remarks We conducted a variability study of 41 Galactic BSGs with well-constrained spectroscopic parameters of Teff and log g (Fraser et al. 2010), using TESS data across sectors 5 − 66. The luminosities of the stars were determined from the modeling of their spectral energy distribution. ●Both pre-processed (PDCSAP) and own extracted/corrected fluxes were explored, taking also into account possible contamination effects in the aperture via assessing the relevant metric. The light curves were de-trended and normalized, and consecutive sectors were joined into larger observing windows. The TESS data of selected BSGs are displayed in Fig. 1. ●We described the time domain by means of three statistical measures: the standard deviation σ, the coherency parameter (ψ2), and the skewness, to assess the degree of scatter, stochasticity, and asymmetry of the TESS data. ●In the frequency domain, we extracted prominent frequencies via iterative pre-whitening and modeled the background signal, so called stochastic low-frequency (SLF) variability, that manifests ubiquitously (Fig. 2). ●Correlations between the variability and stellar parameters were assessed by calculating the Pearson's correlation coefficient r with a p-value significance. ●We found a strong positive correlation between σ and log(L/L⊙ ) (r = 0.74, p < 0.0001); the more luminous BSGs display greater variability amplitude (Fig. 3). The frequencies of the less massive stars comply with their rotational periods, attributing variability to surface spots or a clumped wind. ●Positive correlation is also seen between ψ2 and log(L/L⊙ ) (r = 0.52, p < 0.001); the more luminous variables possess smooth light curves with a higher degree of coherency (yet with larger fluctuations across the different sectors) compared to less luminous ones (Fig. 3). ●The positive trend between Γe and the skewness (r= 0.33, p = 0.04) implies that younger and less massive BSGs display short-term brightening, which is characteristic of wind, line-driven, instabilities (Krtička & Feldmeier 2021). ●Less bound to the above trends are several of the α Cygni variables (outer circles; Fig. 3), a class widely acclaimed as post-RSGs (see also Fig. 4). The irregularity in their TESS data follows the inhomogeneity in the excitation mechanisms of their pulsations, these including oscillatory convection modes and radial strange modes (Saio et al. 2013). ●A significant trend is seen between the SLF variability amplitude α0 and logTeff (r = 0.59, p < 0.0001); the background signal is pronounced at the earlier stages of evolution and drops as the stars age (Fig. 4). ●Our study of post-main-sequence BSGs (Kourniotis et al. 2025) complements studies on early-phase stars (Bowman et al. 2020; Shen et al. 2024) showing that the SLF variability is amplified at or shortly after the end of the main-sequence phase. ●Less plausible is the scenario that SLF variability is generated by subsurface convection zones driven by the iron opacity peak. The latter case raises expectations for an amplified SLF signal towards the upper cool Hertzsprung-Russell (HR) diagram (Cantiello et al. 2021; Shen et al. 2024) that, however, does not comply with the current picture (Fig. 4). ●On the other hand, when confining ourselves to lower luminosities that exclude the α Cygni stars, we find a weak positive trend between α0 and the stellar luminosity (mass). Using Gaia distances we calculated the absolute G-band magnitudes and evaluated them against α0. In Fig. 5, the linear fit to stars fainter than -6.4 mag is depicted with a solid line. ●The latter trend is similar to that of BSGs in the Large Magellanic Cloud (dashed line; Bowman et al. 2019a), implying that SLF variability is possibly driven by a metallicity-independent mechanism, such as core convection. ●A study with a larger statistical sample would enable improved mapping of variability in BSGs across the HR diagram and higher statistical significance. A systematic exploration with TESS of an extended sample of Galactic BSGs using machine learning is ongoing (Kourniotis et al. in prep.). References Fraser et al. 2010, MNRAS, 404, 1306 Kourniotis et al. 2025, A&A, 697, 152 Bowman et al. 2019a, Nat. Astron., 3, 760 Krtička & Feldmeier 2021, A&A, 648, A79 Bowman et al. 2020, A&A, 640, A36 Saio et al. 2013, MNRAS, 433, 1246 Cantiello et al. 2021, ApJ, 915, 112 Shen et al. 2024, ApJS, 275, 2 Ekström et al. 2012, A&A, 537, A146